# View-Through Attribution: When It Lies and When It Helps

Source: https://attributionplaybook.com/blog/2026-10-11-view-through-attribution-guide.html

October 11, 2026

*View-through credit is either insightful or fiction depending on how you use it. The setups where it earns its place, and the ones where it misleads.*

View-through attribution gives credit to ads that were seen but not clicked. It is either a crucial correction to click-only measurement or an elaborate fiction, depending on how it is set up. Most teams use it in the configuration that produces fiction. Here is when view-through earns its place and when it misleads.

## What view-through actually measures

A view-through conversion means someone saw an ad, did not click, and converted later within the lookback window. The charitable interpretation: the ad contributed to awareness that led to the purchase. The skeptical interpretation: the ad was wallpaper and the purchase would have happened anyway.

Both can be true simultaneously. Display and video ads genuinely build awareness, and they also genuinely get credit for conversions they did not cause. The job is to separate the two, not to pick a side.

## When it lies: the configurations that produce fiction

Long lookback windows are the biggest fiction generator. A 30-day view-through window credits a display impression from three weeks ago for a purchase driven by yesterday's branded search. Shorten the window to match your actual consideration cycle; for most ecommerce that is three to seven days, not thirty.

Frequency capping matters too. Without it, heavy users accumulate dozens of view-through touches and the model spreads credit across all of them. And view-through on retargeting audiences is nearly pure fiction: those users already intended to buy, so the ad gets credit for intent it did not create.

## When it helps: the setups that earn their place

View-through is most honest for upper-funnel prospecting with short windows and incrementality testing. Run a holdout: suppress the prospecting audience in one geo and compare conversion lift. If the exposed group converts more, view-through is measuring something real. If not, it is wallpaper.

It also helps for video, where clicks are rare but attention is real. A completed video view followed by a branded search within a few days is a plausible causal chain. Model it, but validate it with the holdout before trusting the numbers.

## A practical view-through policy

Use short windows tied to your sales cycle. Exclude retargeting pools from view-through credit. Require incrementality evidence before letting view-through influence budget. And never let view-through be the metric that justifies its own spend; that is circular.

Report view-through separately from click-through, always. Blended numbers hide the composition, and the composition is the whole story. A channel that is 90 percent view-through credit is a very different bet from one that is 90 percent click credit.

## Make incrementality a habit, not a project

One holdout test is good. A quarterly cadence of them is transformative. Calendar a recurring incrementality check for each major channel: one geo suppressed, one exposed, measured over two to four weeks. The cost is modest and the learning compounds.

Document every test in a shared log: hypothesis, setup, result, decision. After a year you own something most marketing teams never build, an internal evidence base for what actually drives lift. Budget fights get shorter when the answer is written down instead of argued from dashboards.

## Frequently asked questions

### Should small brands use view-through at all?

Only with short windows and a holdout test. Without validation, small budgets get distorted fast by view-through credit assigned to upper-funnel ads that did not cause the sale.

### What lookback window should we start with?

Match it to your consideration cycle. Impulse products: one to three days. Considered purchases: seven to fourteen. Thirty days is the default that produces the most fiction.

### Does view-through double-count with click attribution?

It can, which is why deduplication rules matter. Decide the hierarchy explicitly: typically a click within the window beats a view, and only one touch gets the credit.

## Frequently asked questions

### What is view through attribution?

It credits an ad with a conversion when someone saw the ad, did not click, and bought later through another path. It tries to measure the real influence of display, video, and social ads that people rarely click but often remember.

### Why is view through attribution controversial?

Because the windows and rules decide the result. A long window credits ads for purchases they barely influenced, which is why platforms like it and finance teams distrust it. The honest use is comparison over time with a fixed window, not absolute truth.

### How should stores use view through data?

As one signal among several: compare campaigns against each other with the same window, sanity check against incrementality tests, and never let view through numbers alone justify a budget increase.

[Home](../index.md) / [Blog](index.md)